In recent years, researchers working within the discipline of applied linguistics and beyond have discovered stimulating opportunities to study human interaction, rituals and behaviours online. But with these opportunities come concerns for the human subjects ‘behind’ online data, most notably, their rights to privacy and freedom from harm. Applied linguists working with data from the internet therefore have a responsibility to continually scrutinise and re-evaluate the ethics of their research methods in line with fast-moving technological developments. In the past decade, an increasing number of scholars have advocated a case-based, context-sensitive approach to the issue of privacy, ethics and internet research. But the mechanisms for applying such an approach are often not made explicit, leaving researchers unsure how to proceed ethically in increasingly complex and shifting research contexts. Using a study of the Mumsnet Talk forum as a case, I show how my self-reflexive and linguistic approach, which draws on established traditions in qualitative research, has helped me to understand what many users of this forum see as normal information flows within this setting. I suggest that such a reflexive-linguistic approach is of value to internet researchers working within the applied linguistic discipline and beyond, who wish to make sensitive, informed ethical judgements that minimise the risk of harm to their participants.
The way I tend to feel about [Mumsnet Talk] is that it’s a bit like talking in an open room such as a pub – the conversation is open to anybody to join but you still have an idea who is a part of it, and it’s very different someone standing and joining in (or even standing and participating non verbally) than somebody hiding in the corner and taking notes. The second feels fine, the third feels like eavesdropping.
These are the words of BertieBotts, a user of the online discussion forum of a popular British parenting website, Mumsnet Talk. In this personal communication, she touches on the complex nature of privacy online, especially within a forum that declares itself to be “public”, thus raising an ethical issue that is of interest, concern and debate for many internet users and researchers. Her words problematise the public/private dichotomy that has been central to the ethical decision-making process for many internet researchers (Androutsopoulos 2008; Sveningsson Elm 2009; Thelwall and Wilkinson 2010). The most up-to-date thinking on the subject of ethics, privacy and internet research has tended to move away from this dichotomy, returning instead to core ethical principles, most notably the maxim do no harm, and advocating an approach to ethics that is case-based and attentive to context (Markham and Buchanan 2012, 2015).
Nissenbaum’s (2010) framework of “contextual integrity” can help internet researchers to move away from the potentially reductive public/private binary. Contextual integrity demands respect for “informational norms”, which Nissenbaum (2010: 141) defines as the “distinctive set of rules governing information flows” in specific contexts. Informational norms, Nissenbaum suggests, are what determine the contextual appropriateness of particular behaviours. She considers, as an example, the appropriateness of sharing information about the condition of our bodies, which she suggests would be expected and usual in a healthcare setting, but probably unexpected and inappropriate in the workplace. Nissenbaum’s (2010) framework demands scrutiny not only of the “norms” of the context in which information is produced, but also those of the context in which it will be re-produced. As Nissenbaum (2010: 142) puts it, “usually, when we mind that information about us is shared, we mind not simply that it is being shared but that it is shared in the wrong ways and with inappropriate others”. In this article, I suggest that paying attention to informational norms facilitates a case-based, context-sensitive approach to internet research ethics. In turn, such an approach can lead to insights about how harm may be caused through violation of these norms, and how such harm can be avoided.
An issue Nissenbaum (2010: 44) does not address in great depth is how researchers and practitioners should go about identifying informational norms: she claims to rely on “an intuitive sense” to understand contexts and types of information. Such reluctance to offer specific guidelines or methods for making ethical judgements is common and understandable amidst the complexities of internet research ethics (Ess 2009). Where specific guidelines are offered, they are likely to break down when applied to online contexts in which multiple, usually distinct audiences are often “collapsed” (boyd 2011; Marwick and boyd 2011). Nevertheless, I would argue that specific methods can be deployed to unpack some of the informational norms that emerge within online research sites or communities, as well as at specific moments of interaction. This paper offers a framework for applied linguists and other internet researchers who wish to do just that. Using a case study of Mumsnet Talk, I detail some of the grounded, ethnographic, self-reflexive and linguistic methods I have employed to identify some informational norms of interaction within this forum. The methods I explicate have been widely used by scholars in a range of contexts, including internet research in the applied linguistics discipline and beyond. However, there is relatively little research that focuses on how these (or indeed any other) specific methods can help researchers to better understand the informational norms that emerge in online research contexts.
Many internet researchers have responded to the perceived over-simplification that results from categorising online contexts as either public or private by acknowledging the “grey areas” between the two poles (Barton and Lee 2013; Ess 2007). Some have conceptualised public and private as opposite ends of a continuum (McKee and Porter 2009; Sveningsson Elm 2009), with intermediate states such as “publicly private” and “privately public” (Lange 2008: 362), or “semi-public” and “semi-private” (Sveningsson Elm 2009: 75) occupying these grey areas. Arguably, however, conceptualising public and private as a continuum prompts the same question, “is this public or private?” though in the more relative form “how public or private is it?” Thus, a continuum can ultimately work to reinforce the same dichotomy (Markham and Buchanan 2015).
Rather than trying to force online contexts into the familiar but often reductive categories of public and private, or somewhere on a sliding scale between the two, many scholars have instead turned their attention to the notion of what is appropriate in context. Danet et al. (1997), Lüders (2015) and Rosenberg (2010), for example, consider participants’ own perceptions and expectations about privacy, audience and information sharing as part of the ethical decision-making process. These authors find that users of various internet sites often have expectations that conflict with notions of public accessibility, as when individuals “maintain strong expectations of privacy” even where the space they use is accessible to all (Lüders 2015: 81). Marwick and boyd (2014: 44) also show how internet users can have a particular audience in mind, regardless of the availability or accessibility of the information they share. One such scenario is hypothesised by BertieBotts in the words that open this paper. BertieBotts recognises that Mumsnet threads are, in theory, “open” to all. However, she expresses very clearly her feeling that a researcher “sitting in the corner taking notes”, so to speak, may not feel appropriate to regular users of the forum. boyd’s (2011: 44) term “intended public” captures the notion that, despite the way contexts frequently collapse in online spaces, so that potentially distinct audiences merge, internet users will often have quite a specific “imagined audience” to whom they will tailor their practices. These users may expect, as BertieBotts seem to, that their audience will pick up on cues that indicate such subtle distinctions.
Nissenbaum’s (2010) framework of contextual integrity places the onus on the researcher to tease out what is normal, expected and appropriate use of data in context. Such a focus on appropriateness is flexible and responsive to emergent and developing forms of social interaction and to the roles users themselves play in the construction of informational norms. But Nissenbaum’s (2010) framework is not without its constraints and limitations. It acknowledges a much wider, indeed potentially infinite range of social contexts than other models of privacy, but it fails to emphasise the potential overlap between contexts (Marwick and boyd 2014). Thus, contextual integrity may carry weight in ritualistic encounters such as a doctor’s appointment, purchasing goods at a supermarket or being interviewed for a job, but it is difficult to maintain the notion of distinct informational norms in relation to the shifting and collapsed contexts that are increasingly typical of online interactions (boyd 2011; Marwick and boyd 2011). A related limitation of Nissenbaum’s framework is her emphasis on context over user: she assumes that all individuals interacting in a particular context will uphold the same informational norms. She makes her argument for the primacy of context with reference to a controlled study in which, she notes:
Individual variability was overshadowed by striking similarities in the degree to which information types and recipient roles were predictive of the respondents’ level of comfort in sharing information. This should put to rest the frequent insinuation that privacy preferences are personal and idiosyncratic.
(Olson, Grudin and Horvitz 2005, in Nissenbaum 2010: 151)
This claim disregards statistically small differences of perception, rendering them irrelevant or anomalous. By discounting the significance of ‘idiosyncrasies’, Nissenbaum fails to acknowledge the potential relevance of individual users’ values and perceptions, and the infinitely variable range of factors they may bring to different contexts, such as past experiences, membership of other communities and orientation to particular groups or categories.
In response to the limitations of Nissenbaum’s (2010) framework, Marwick and boyd (2014) suggest that informational norms are not fixed, but negotiated around a particular social situation at a particular time: that they are achieved by users. I make the related point that, whilst generalised norms can often be identified within a particular community or research site, they can only serve as a guide; as patterns of expectation, rather than a fixed set of rules. Taking account of these important qualifications, I present a slightly adapted definition of informational norms as identifiable patterns of expectation, achieved in social context, about the normal and appropriate use of data.  The discussion in Sections 3 and 4 will show what kind of insights can be revealed through close attention to internet users’ multiple perceptions and expectations about their information sharing practices.
In this section, I outline a framework for identifying informational norms in online contexts with reference to a case study of Mumsnet Talk. The Mumsnet website was founded in the year 2000 by the British entrepreneur Justine Roberts, with the aim of “mak[ing] parents’ lives easier by pooling knowledge, advice and support” (Mumsnet Limited 2015). The site includes relatively static, informative pages on topics such as money, work and education, a reviews section, bloggers network, and the discussion board for which it is most well-known: Mumsnet Talk. This board offers one primary mode of interaction: contributions to threads, which are initiated minute-by-minute at busy times. The forum is defined on the site itself as “public”, as the screen shot from the Talk home page (Figure 1) clearly shows. Although members have to sign up with an email address and username to contribute to Mumsnet Talk, the threads posted to this forum are accessible to anyone with an internet connection. Talk threads often appear in internet search results and, unlike social network sites such as Facebook, there are limited options for users to control the accessibility of their posts.
My engagement with Mumsnet Talk leads me to conceptualise this forum as a type of participatory culture (Jenkins 2009), in that it has low barriers to participation and, as noted above, its users share knowledge, advice and support. To some degree, Mumsnet Talk may be conceptualised as an affinity space (Gee 2004), in that its contributors are brought together by a shared interest in parenting. However, I would argue that Gee’s (2004: 67) description of affinity spaces as those where “people affiliate with others based primarily on shared activities, interests, and goals, not shared race, class culture, ethnicity, or gender” is not consistent with the nature of Mumsnet Talk. Both my own study and the work of Pedersen and Smithson (2013) suggest that Mumsnet users are relatively homogeneous in terms of identity categories such as gender, sexuality and class. Its users, by and large, identify as female parents, and demographic data collected for Pedersen and Smithson’s (2013) study, as well as the 2009 Mumsnet census, suggests that they are predominantly working mothers, that most have a household income above the national average and that a large number have university degrees. Similarly, in my own study, I have found that contributors persistently position themselves as female, heterosexual, middle-class parents. In this paper, I refer to Mumsnet Talk as both a community and an in-group. These terms foreground my sense that this is a space in which members feel some sense of belonging, rather than a space for fleeting interactions around a shared interest. The observations, analyses and reflections that are presented in this paper support this claim.
My case study draws on data collected from Mumsnet Talk in an exploration of the discursive construction of motherhood through digital interaction. As part of this study, I selected fifty threads that share the common themes of describing, enacting or negotiating gendered and parental identities. From these fifty threads, I went on to choose two threads that contrast in style, yet both foreground participants’ gendered identities, for further microlinguistic and discursive analysis. In the first thread, Your identity as mother, Mumsnet users scrutinise their perception of themselves as mothers. In the second thread, Can we have a child exchange?, contributors enter in to a mock bidding war, introducing, describing and offering their children for exchange in the style of a classified advertisement. 
I take a grounded approach to this study, allowing insights to emerge through the process of data construction and analysis. This approach draws on established traditions in qualitative research, most notably ethnography and grounded theory. It is supported by a self-reflexive stance that involves constant scrutiny and re-evaluation of my methods, my position as a researcher and my relationship with participants. Below, I outline five specific methods employed as part of what I call this reflexive-linguistic approach:
Adopting a participant stance;
Engagement with participants and gatekeepers, and
These five methods represent tools in the applied linguist’s repertoire that have facilitated my understanding of the informational norms in my research context and can help others to do the same. Although they are numbered chronologically, they do not necessarily represent distinct steps or a staged process. Rather, internet researchers may adopt just one method, or all five, as a comprehensive toolkit, as they strive to identify the informational norms of their research contexts and ultimately to anticipate and mitigate potential causes of harm.
The process of data construction for the Mumsnet study has involved “systematic observation” (Androutsopoulos 2008) of the Talk forum over a period of five months in 2014. During this observational period, I visited the forum regularly and kept detailed notes. I observed the discussions that were taking place but did not contribute to them. My early observations of the Talk forum led me to assume that there was little likelihood of my research causing any kind of harm through data exposure or violation of privacy because I perceived all of the data I would be using to be in the public domain. I therefore did not feel the need to take steps such as asking potential participants for their informed consent. However, through systematic observations of the site, I began to recognise that this community of users, many of whom have been regular contributors since its inception, have established norms of participation and interaction that are much more complex and subtle than the “public” label would suggest. My developing recognition that Mumsnet users rarely expect a general and infinitely varied public to engage with their posts, together with my growing ability to understand the site from the perspective of a participant, brought me to the realisation that my research did have the potential to cause harm, because it might violate users’ expectations about normal or appropriate use of their interactions on the forum.
I have used memos throughout the data construction process to document my observations, personal responses and data selections. These memos are a form of electronic record usually associated with grounded theory (Charmaz 2014; Corbin and Strauss 2008; Glaser and Strauss 1967), though they are also akin to the field notes associated with ethnographic research. The notes I write tend to be quite detailed and sometimes include in-depth and analytical comments. This style of note taking is supported by the “chrono-malleable” (Markham 2004: 103) nature of digital interactions, meaning, for example, that I can leave the Mumsnet webpage, write a memo, then immediately return to the same page. I can also write memos without leaving or disturbing the setting, in a separate window or on paper, my eye and thoughts moving intermittently between the two.
The process of memo writing is central to my self-reflexive approach. It provides an opportunity to bring myself in to the analysis by reflexively examining my own preconceptions and my role as a “constructor” of data (Charmaz 2014; Mason 2002). I use memos to record my personal feelings, impressions and developing interpretations, and to reflect on any factors that I believe might influence these interpretations. It is often through the process of memo writing that I consolidate my understanding of the way Mumsnet users interact within the forum and its value in their lives. For example, I have used memos to document and reflect upon Mumsnet users’ comments about and perceptions of anonymity, including the strategies they employ to achieve anonymity and the benefits of being anonymous. My reflections on this topic led me to recognise what anonymity can mean to users of this forum, its value, some of the ways in which anonymity could be breached, and how such a breach might cause harm.
The value of researcher participation is well documented in ethnographic internet research. Scholars such as Baym (1993), Pederson and Smithson (2013) and Rosenberg (2010), for example, have all gained unique insights into the online communities they study through their participatory roles. However, early in the development of my project, I did not feel able to position myself as a participant. Although I am a mother, I have never contributed to the Mumsnet Talk forum and became a member only after deciding to undertake this research. For these reasons, restricting myself to a wholly observational role initially felt personally, practically and ethically appropriate. During the process of data construction, however, my stance and my relationship to Mumsnet users began to shift. I began to understand that, by positioning myself as an observer, I was failing to acknowledge my growing affinity with Mumsnet users and my very real engagement with the site. I realised it was almost impossible to neatly separate my roles as mother and researcher and that in many ways I actually felt very much a part of the Mumsnet community. These reflections led to an adjustment in my self-positioning, from observer to observer-participant, an intermediary position outlined by Mason (2002; see also Gold 1958, who uses the term observer-as-participant).
By situating myself within the research site, I was better able to understand it as a participant, even though I remained a ‘silent’ user, as I did not contribute to threads. For example, I was able to question whether I was an appropriate audience to be reading, capturing and analysing Mumsnet threads. I started to ask questions such as “how would I feel if I’d had a bad day and been negative about my child or my husband, or shared an intimate personal experience, and those posts were printed in a publication without my knowledge?” As a result, I began to recognise certain informational norms. For example, many contributors often seemed to address quite a specific “intended public” (boyd 2011) that I felt did not include me (as a researcher) and most would not expect a researcher to take an interest in their contributions to a busy forum. When I adjusted my stance and tried to understand the forum from the perspective of a participant, I was able to recognise the potential for my research to cause harm through violation of such norms and to re-evaluate my ethical choices accordingly. One of the most significant changes I made as a result of these considerations was to contact all of the Mumsnet users whose words I wished to quote and/or analyse in detail, asking for their informed consent, giving them the option to have their usernames anonymised and making it clear that they were free to withdraw at any time. Those who explicitly declined to consent, or did not respond, have not been included in my study. Their posts and usernames were removed from threads at an early stage, together with any reference made by other participants to these users or their posts. Allowing participants to self-select has limited the diversity of voices that are heard in my study and potentially affects my analysis of interaction, because I cannot analyse threads in full. However, because Mumsnet threads do not tend to unfold in chronological sequence and participants often do not respond directly to one another, this is considered a reasonable adjustment. Both of the threads I analyse remain coherent overall and several interactional sequences, where participants do interact directly with one another, can still be identified, as some of the examples offered in Section 4.2 will show.
In collaboration with Mumsnet staff (see below), I contacted individual users through the Mumsnet private messaging system. In my message to potential participants, I offered a brief summary of the aims of my research and provided a link to my personal blog, which included detailed information about me and my study on a dedicated page, as well as an extensive posting history that served to chart the development of my research over a period of approximately eighteen months.  In both my message to potential participants and my blog, I introduced myself as a researcher and also “a mother to a 4 and 5 year old”. By doing so, I positioned myself as both a parent and a woman. These aspects of my own identity had a significant impact on the research process, as noted above. I felt that the similarities between myself and many other Mumsnet users would also influence the way members responded to me, and in particular their trust of me as a researcher (see Section 4.3).
Before making any contact with Mumsnet users themselves, and indeed, before even embarking on my study, I contacted Mumsnet staff by email, explaining what I wanted to do and asking for their advice about whether and how I could use data from the site. I saw Mumsnet staff as gatekeepers – as officials with a good knowledge of the site and its regulations, who are able to control the flow of information on the site and permit or restrict access to it. With very little knowledge of Mumsnet at the time, I felt it was important to seek these gatekeepers’ support and guidance with regard to what would constitute appropriate (and also legal) use of content from Mumsnet Talk. Mumsnet staff suggested early on that I contact users whose words I wished to quote, and that the most straightforward way to do so would be through the Mumsnet private messaging system.
After contacting the Mumsnet users whose words I wished to analyse in detail and/or quote, I was surprised by how many were keen to engage in discussion about my research, my ethical approach and their perceptions of Mumsnet Talk in general. One user who expressed particular interest in my research was BertieBotts, a long-standing member of Mumsnet and one of only two users who contributed to both of the threads I selected for close analysis. After BertieBotts replied that she “really appreciate[d] being asked and responded to in this way”, I asked her to elaborate on what it was she had liked about my approach. The conversation that ensued further shaped my understanding of informational norms in this context; namely, that contributors’ intended public is generally other Mumsnet users (and sometimes a very specific group of users – see Section 4.2) and that outsiders – people who are not regular users of the forum – are unlikely to understand the idiosyncrasies of this online community (see Section 4.3). Indeed, my understanding of all the informational norms I had identified by this stage was strengthened and deepened by engagement with Mumsnet users. The variety of responses I received to my initial request for consent revealed a further truth: that I would never be able to identify absolute, universal rules; to predict every individual Mumsnet user’s perceptions and expectations surrounding their participation within the Mumsnet Talk forum (see Section 4.4).
My understanding of informational norms, and in particular, the strategies employed by Mumsnet users to achieve a level of privacy, continued to develop through the process of linguistic analysis in my study. I have taken a grounded, inductive approach to linguistic analysis, so that the tools I adopt and the particular linguistic features upon which I focus are influenced by the context of my study, my aims, and by the nature of the threads I analyse. The work of Du Bois (2007, 2014) has been particularly relevant to my identification of informational norms in Mumsnet Talk. Du Bois’ focus on intersubjective alignment (2007) and dialogic resonance (2014) has facilitated my exploration of the way Mumsnet users achieve a degree of privacy through their positioning of self in relation to others. For example, I explore the way participants position themselves as part of an in-group of Mumsnet users, and also align with smaller groups and specific individuals, through their adoption of shared linguistic and digital resources (see Section 4.2). Such analyses have allowed me to scrutinise who users’ intended publics may be and to pinpoint moments at which those intended publics can shift, for example from a relatively general audience of Mumsnet users to a more specific group with a shared interest or connection.
This final method is particularly relevant for applied linguists, who can draw on their expertise to identify and find evidence of informational norms through close scrutiny of linguistic practices.
The application of a reflexive-linguistic approach to the Mumsnet study has led me to identify four broad informational norms within Mumsnet Talk:
“Anonymity is an extremely strong value”;
Contributors’ intended public is generally other Mumsnet users;
‘Outsiders’ are likely to misunderstand the community, and
Individual users have variable expectations about information sharing.
These informational norms represent patterns of expectation about the normal and appropriate use of data from this forum, except for the fourth, which draws attention to the potential variability of Mumsnet users’ expectations about information sharing within and beyond the forum. Each norm will be outlined in turn below.
The title of this informational norm is taken from the words of freespirit, a contributor with whom I discussed the ethical implications of my research in depth via private message. In addition to the direct discussion with freespirit, the importance of anonymity and the strategies Mumsnet users employ to achieve it became apparent quite quickly through my observations of the site. For example, the practice of adopting pseudonymous usernames offers a degree of anonymity in an open forum. Further, users often engage in what I call selective sharing, rarely disclosing personal or identifying information such as images, names or places or, as I later discovered through direct communication with participants, altering details about their lives, such as the gender by which they refer to their child. Through their adoption of these practices, Mumsnet users can be said to achieve a degree of privacy without compromising their participation in this highly accessible forum. They are able to take control of what is visible to others, keeping certain information from view.
Mumsnet users’ concern to remain anonymous was also apparent at the point of contact with participants. Almost all users who responded to my message wanted to see their contributions before giving their consent, with many making it clear that they wanted to check they hadn’t written anything that could make them identifiable beyond the Mumsnet community. By taking an observer-participant stance, I was able to recognise that participants may feel their privacy has been breached if a post they consider makes them identifiable (but which I might not recognise as such) is shared beyond this community. This is particularly true of posts to the “chat” section of the forum, from which my selected threads have been chosen, and which disappear from the site after 90 days. By saving and reproducing these posts, however, I make them available to quite a different audience for a much longer period of time.
As well as adopting pseudonyms, some users regularly change their usernames, particularly if they feel they have revealed something that could make them identifiable beyond the Mumsnet community. Selected excerpts from the forum in Extract 1 illustrate this point.
The excerpts included in extract 1 show that some Mumsnet users see themselves as having a double layer of anonymity, whereby they are unidentifiable both within and outside of the Mumsnet community. The fact that many of my participants requested to have their usernames further anonymised for the purposes of my study supports this point. The concept of anonymity, accordingly, is not conceptualised here in terms of a sharp distinction between ‘online’ and ‘offline’ identities. Rather, I suggest that preserving anonymity is a complex matter of protecting individuals’ sense of dignity and privacy, and respecting the steps they take to control their self-presentation in particular contexts.
This informational norm is neatly expressed by BertieBotts in the quotation that opens this paper, where she suggests that she has a sense of who is part of a conversation, even if they do not participate verbally. My impression that Mumsnet users generally write for quite specific intended publics is further evidenced by close scrutiny of the linguistic and digital resources Mumsnet users draw on in their interactions. For example, through my analysis of the Your identity as a mother thread, I note that contributors consolidate an in-group identity of Mumsnet users by employing resources that are ubiquitous in Mumsnet Talk, some of which are listed below.
Acronyms and abbreviations: ‘DD’ (darling daughter), ‘MN’ (Mumsnet), ‘IYSWIM’ (if you see what I mean)
Strikethrough text: ‘don’t fight it
or are you shallow’
A range of brackets or asterisks to represent non-verbal actions, sounds or ‘stage directions’: ‘<sideways look and inward tut tut>’
Capitals for emphasis/volume: ‘I am right, I AM!’
‘Tagging’ another user (which automatically notifies the user they have been ‘mentioned’): ‘Nonie don’t take this the wrong way’
Nonstandard vocabulary/neologisms: ‘mumming’, ‘teenagerdom’, ‘Daily Fail’
Creative use of punctuation: ‘I have a name!!!! I am a person!!’
Participants’ adoption of such shared linguistic and digital resources can be said to invoke a specific intended public of Mumsnet users who have ‘insider’ knowledge. Some of these resources, particularly acronyms and abbreviations, strikethrough text and nonstandard vocabulary, restrict access to meaning, making it difficult for readers who are not familiar with Mumsnet Talk to interpret elements of posts to this forum. Such markers of in-group membership can be said to add another layer of privacy to Mumsnet interactions, along with pseudonymous usernames, name-changing practices and selective sharing. By using these resources, Mumsnet users can be said to hide content “in plain sight” (Marwick and boyd 2014: 8).
Further observation and analysis of Mumsnet Talk suggests that Mumsnet users not only position themselves within a wider in-group, but also at times within smaller, even more exclusive in-groups of specific contributors. For example, I have identified moments at which smaller groups of users can be seen to break away from a larger party, akin, to use BertieBotts’ analogy, to a conversation between friends in “an open room such as a pub”. This can be seen in some extended threads, in which a core group of contributors emerges and interactions increasingly revolve around shared knowledge and understanding between contributing individuals, rather than a particular topic or theme. People joining these threads tend to affirm the in-group status of core contributors by asking permission to join or enquiring about the “rules” of the thread. Contributors therefore increasingly address quite a restricted in-group of Mumsnet users.
My linguistic analysis also shows how even smaller, more ephemeral in-groups can be constructed around a particular topic or connection within a more general thread. Such closed-group interactions can be identified in Extract 2, a sequence  from Your identity as a mother in which the contributors Thurlow, Viglioso and Dysfunctional address the overall question of the thread – what is your identity as a mother – with a focus on how parental identities can be defined by particular schools of thought such as “attachment parenting”. At moments like these, Mumsnet users’ intended public becomes even more specific, being restricted, even, to a few named users.
As well as addressing other users quite explicitly in the opening lines of each post, these contributors construct and affiliate with a more restricted in-group at this moment through the selective reproduction (Du Bois 2014) of linguistic elements from one another’s posts. For example, in post 59, Viglioso’s use of the pronoun “some” (lines 5 and 10) echoes generic, impersonal naming devices used by Thurlow in post 39 such as “someone” (line 5) and “some people” (line 8). Both contributors also draw attention to the way these others “define themselves” (Thurlow, lines 8–9; Viglioso, line 6).
Further evidence of a narrowed intended public in this sequence can be found in post 80. Here, Dysfunctional’s alignment with Thurlow can be identified in her statement “I know what you mean by the AP ‘tribe’” (line 1), in which Dysfunctional directly addresses Thurlow with the pronoun ‘you’ and reproduces what she takes as Thurlow’s category “tribe” (though in fact, it is Viglioso who introduces this category). Dysfunctional’s use of inverted commas here marks her first use of this category as an overt reproduction of the words of another. She uses it for a second time, however, without speech marks (line 4), suggesting that this category is now naturalised to her own repertoire; that she has adopted Thurlow’s words as her own. Again, the linguistic resonances between these posts builds on the construction of a more restricted in-group. As a result, this sequence sees a shift from a relatively general intended public of Mumsnet users to a more intimate and specific audience that is constructed around a particular interactional moment.
BertieBotts succinctly expresses the potential for harm, should the norm “Mumsnet users’ intended public is generally other Mumsnet users” be flouted, when she writes in a private message that “people understandably feel a bit violated when they have felt they were talking somewhat privately, even though it is an open forum”. Her words make it clear that she is aware of the openness of the forum, but that the accessibility of the forum is often at the back of her mind. My identification of this informational norm again leads me to judge that it would not be ethical to reproduce data from Mumsnet Talk without the prior knowledge and consent of those who produced it, probably with quite a different intended audience than myself in mind.
It follows from the previous informational norm that Mumsnet users do not necessarily expect individuals outside of their intended public to understand the spirit of their threads and individual posts. This is not a simple matter of assuming that others will misunderstand acronyms or other in-group resources, but a fear that someone beyond the intended public could co-opt and potentially misrepresent their words. BertieBotts summarises this fear in quite emotive terms in a private message to the researcher (see Extract 3). She makes it very clear that she sees Mumsnet as a “community” to which feels a sense of belonging and that others’ misunderstanding of this group would feel like a personal insult, and could cause distress.
On further discussion (see Extract 4), BertieBotts revealed what it was that helped her to overcome her mistrust of me as a potential outsider.
What becomes clear from these comments is that BertieBotts’ decision to give her informed consent had a lot to do with what would happen to her words once I adopted them for a new purpose; she wanted to know who was going to use her words, as well as how and why they would be used. BertieBotts’ fears were allayed by her sense that I had much in common with her, and other Mumsnet users, in terms of my political views, my status as a mother (and presumably also as a woman) and even my writing style. She explains and justifies her agreement by negotiating my position as a sort of honorary member of the community, by virtue of these shared traits. As such, I am trusted with her information because I am seen as part of her intended public. BertieBotts’ message also draws attention to the importance of my participatory stance. Had I not embraced my involvement with the Mumsnet community, my message and blog post, as well as subsequent analyses, would have very likely been less sensitive to Mumsnet users’ expectations and concerns. Embracing my participatory role made me a more trustworthy researcher.
The fear of being misunderstood by outsiders could explain the resistance I encountered from some potential participants when first asking for their informed consent. When I contacted contributors to the Can we have a child exchange? thread, for example, their responses were often quite tentative. Many wanted to know precisely how their posts would be used and expressed disbelief that they could be useful for my research project or confusion as to why I wanted to use them. These contributors’ reluctance to participate in my study points to a perceived mismatch between the spirit in which the thread was written and my research topic. I believe some contributors were afraid that their words might be taken out of context and interpreted in a way that would be incongruous with their perceived intentions and the humorous, ironic tone of the thread, and in the worst case, that they might be vilified for their words. Such a situation would very likely lead to distress, should participants learn how they had been represented.
Although I have identified three informational norms that can be broadly applied to Mumsnet Talk, my experiences also lead me to conclude that the only way to know how every individual participant will feel about their words being used for research purposes is by asking them. Even then, participants may not be able to fully realise or express potential reservations, or their feelings could change over time. The variability of participants’ responses to my requests for informed consent supports my identification of this informational norm. Some contributors to the same thread consented enthusiastically, and some even thanked me for asking them, noting that they had assumed their words were open to all. Others emphatically refused to consent and seemed offended by my request. These contributors often offered no particular justification for their refusal and their reasons could not always be deduced from the content of their posts or the theme of the overall thread to which they contributed. I therefore suggest again, contrary to Nissenbaum’s (2010) view, that informational norms should be seen as patterns of expectation, not rules. These patterns can be identified by adopting the approach outlined in this paper, which can assist the researcher in making informed, sensitive, ethical judgements. However, there is no guarantee that these patterns can be applied to all users of a given site or contributors to a particular interaction. What is far more likely is that informational norms will always be subject to a degree of idiosyncratic variability.
This paper shows that a reflexive-linguistic approach can lead to a sophisticated and nuanced understanding of informational norms in an online research site, which in turn can facilitate sensitive and well-considered ethical judgements. Some valuable insights in relation to informational norms within the Mumsnet community, which can be applied to research in similar contexts, emerge as a result of this approach. In particular, I show that Mumsnet users are able to exercise autonomy and agency in imaginative ways to control and shape the accessibility of their posts, their intended publics and the degree to which they are identifiable as single users. By doing so, Mumsnet users are able to maintain a degree of privacy and anonymity in a highly accessible space at specific moments of interaction. I also show that the concept of anonymity is complex, and not just a matter of internet users being unidentifiable ‘offline’. Rather, in spaces such as Mumsnet Talk, where contributors use pseudonyms, anonymity is also a matter of protecting their sense of privacy and dignity within this context.
Methodological approaches to internet research ethics remain under-researched. I propose that there is a need for further work around the strategies scholars can use to identify informational norms within internet sites and moment-by-moment online interactions. Ultimately, further investigation of approaches to internet research ethics will support researchers in making sound ethical judgements and minimising the risk of harm to their participants.
My thanks first go to the Mumsnet users who allowed me to reproduce their posts, originally written with quite a different audience in mind, in my study as a whole and in this paper. Thanks especially to BertieBotts and freespirit, who also allowed me to draw on our personal correspondence, and whose generous engagement with my research cannot be underestimated in the development of my approach to internet research ethics. Thanks also to the editors of this special issue, to two anonymous reviewers, and to Judith Baxter, who all offered incisive and helpful comments on earlier versions of this paper. Any shortcomings that remain are entirely my own.
Androutsopoulos, Jannis. 2008. Potentials and limitations of discourse-centred online ethnography. Language@Internet, 5, article 9. http://www.languageatinternet.org/articles/2008/1610 Search in Google Scholar
Barton, David & Carmen Lee. 2013. Language online: Investigating digital texts and practices. London & New York: Routledge. Search in Google Scholar
Baym, Nancy. 1993. Interpreting soap operas and creating community: Inside a computer-mediated fan culture. Journal of Folklore Research 30(2/3). 143–176. Search in Google Scholar
boyd, danah. 2011. Social network sites as networked publics: Affordances, dynamics, and implications. In Zizi Papacharissi (ed.), A networked self: Identity, community and culture on social network sites, 39–58. New York & London: Routledge. Search in Google Scholar
Charmaz, Kathy. 2014. Constructing grounded theory, 2nd edn. Los Angeles, London, New Delhi, Singapore & Washington, DC: Sage Publications. Search in Google Scholar
Corbin, Juliet & Anselm Strauss. 2008. Basics of qualitative research: Techniques and procedures for developing grounded theory, 3rd edn. London, California, New Delhi, Singapore: Sage Publications. Search in Google Scholar
Danet, Brenda, Lucia Ruedenberg-Wright & Yehudit Rosenbaum-Tamari. 1997. “Hmmm… Where’s that smoke coming from?” Writing, play and performance on Internet Relay Chat. Journal of Computer-Mediated Communication 2(4). doi:10.1111/j.1083-6101.1997.tb00195.x Search in Google Scholar
Du Bois, John W. 2007. The stance triangle. In Robert Englebretson (ed.), Stancetaking in discourse: Subjectivity, evaluation, interaction, 139–182. Amsterdam: Benjamins. Search in Google Scholar
Du Bois, John W. 2014. Towards a dialogic syntax. Cognitive Linguistics 25(3). 359–410. Search in Google Scholar
Ess, Charles. 2007. Internet research ethics. In Adam Joinson, Katelyn Mckenna, Tom Postmes & Ulf-Dietrich Reips (eds.), The Oxford handbook of internet psychology, 487–502. New York: Oxford University Press. Search in Google Scholar
Ess, Charles. 2009. Foreword. In Heidi A. McKee & James E. Porter (eds.), The ethics of internet research: A rhetorical, case-based process, xiii–xvi. New York: Peter Lang Publishing Inc. Search in Google Scholar
Gee, James Paul. 2004. Situated language and learning: A critique of traditional schooling. New York & London: Routledge. Search in Google Scholar
Glaser, Barney G. & Anselm L. Strauss. 1967. The discovery of grounded theory: Strategies for qualitative research. Chicago: Aldine. Search in Google Scholar
Gold, Raymond L. 1958. Roles in sociological field observations. Social Forces 36(3). 217–223. Search in Google Scholar
Jenkins, Henry. 2009. Confronting the challenges of participatory culture: Media education for the 21st century. Cambridge, MA & London, England: The Massachusetts Institute of Technology Press. Search in Google Scholar
Lange, Patricia G. 2008. Publicly private and privately public: Social networking on YouTube. Journal of Computer-Mediated Communication 13. 361–380. Search in Google Scholar
Lüders, Marika. 2015. Researching social media: Confidentiality, anonymity and reconstructing online practices. In H. Fossheim & H. Ingierd (eds.), Internet research ethics, 77–97. Norway: Cappelen Damm Akademisk. Search in Google Scholar
Markham, Annette. 2004. Internet communication as a tool for qualitative research. In David Silverman (ed.), Qualitative research: Theory, method and practice, 95–124. London, Thousand Oaks, New Delhi: Sage. Search in Google Scholar
Markham, Annette. 2013. Undermining “data”: A critical examination of a core term in scientific inquiry. First Monday 18(10). http://firstmonday.org/article/view/4868/3749 Search in Google Scholar
Markham, Annette & Elizabeth Buchanan. 2015. Internet research: Ethical concerns. In James D. Wright (ed.), International encyclopedia of the social and behavioral sciences. Elsevier. Search in Google Scholar
Markham, Annette, Elizabeth Buchanan & The AoIR (Association of Internet Researchers) Ethics Working Committee. 2012. Ethical decision-making and Internet research: Recommendations from the AoIR ethics working committee (Version 2.0). http://aoir.org/reports/ethics2.pdf Search in Google Scholar
Marwick, Alice & danah boyd. 2011. I tweet honestly, I tweet passionately: Twitter users, context collapse, and the imagined audience. New Media & Society 13(1). 114–133. Search in Google Scholar
Marwick, Alice & danah boyd. 2014. Networked privacy: How teenagers negotiate context in social media. New Media & Society 16(7). 1051–1067. doi:10.1177/1461444814543995 Search in Google Scholar
Mason, Jennifer. 2002. Qualitative researching, 2nd edn. London, Thousand Oaks, New Delhi: Sage Publications. Search in Google Scholar
McKee, Heidi & James E. Porter. 2009. The ethics of internet research: A rhetorical, case-based process. New York: Peter Lang Publishing Inc. Search in Google Scholar
Mumsnet Limited. 2015. About us. http://www.mumsnet.com/ Search in Google Scholar
Nissenbaum, Helen. 2010. Privacy in context: Technology, policy, and the integrity of social life. Stanford, CA: Stanford University Press. Search in Google Scholar
Pedersen, Sarah & Janet Smithson. 2013. Mothers with attitude – How the Mumsnet parenting forum offers space for new forms of femininity to emerge online. Women’s Studies International Forum 38. 97–106. doi:10.1016/j.wsif.2013.03.004 Search in Google Scholar
Rosenberg, Åsa. 2010. Virtual world research ethics and the private/public distinction. International Journal of Internet Research Ethics 3. 23–37. Search in Google Scholar
Sveningsson Elm, Malin. 2009. How do various notions of privacy influence decisions in qualitative internet research? In Annette Markham & Nancy Baym (eds.), Internet inquiry: Conversations about method, 69–87. Los Angeles, London, New Delhi, Singapore & Washington, DC: Sage. Search in Google Scholar
Thelwall, Mike & David Wilkinson. 2010. Researching personal information on the public web: Methods and ethics. Social Science Computer Review 29(4). 387–401. Search in Google Scholar
© 2017 Walter de Gruyter GmbH, Berlin/Boston